NELLCO

NELLCO Legal Scholarship Repository

New York University Public Law and Legal Theory
New York University School of Law
Working Papers

2-1-2013

Toleration: Is There a Paradox?

Jeremy J. Waldron
NYU School of Law, jeremy.waldron@nyu.edu

Follow this and additional works at: http://lsr.nellco.org/nyu_plltwp

Recommended Citation
Waldron, Jeremy J., "Toleration: Is There a Paradox?" (2013). New York University Public Law and Legal Theory Working Papers. Paper
376.
http://lsr.nellco.org/nyu_plltwp/376

This Article is brought to you for free and open access by the New York University School of Law at NELLCO Legal Scholarship Repository. It has been
accepted for inclusion in New York University Public Law and Legal Theory Working Papers by an authorized administrator of NELLCO Legal
Scholarship Repository. For more information, please contact tracy.thompson@nellco.org.
Toleration, Supererogation, & Moral DutiesMay 31-June 1, 2012
Conference in Honour of David Heyd Hebrew University of Jerusalem

TOLERATION: IS THERE A PARADOX?

Jeremy Waldron

I
Is there a paradox involved in toleration? Many philosophers think there is.
Though they dont all express it the same way, a number of the distinguished
authors in David Heyds collection, Toleration: An Elusive Virtue say that the very
idea of toleration, properly understood, involves something approaching a
pragmatic if not a logical contradiction.1 We tolerate a practice when we judge it
to be wrong but nevertheless refrain from acting against itin ways that would
normally be thought appropriate for things of which such an adverse judgment has
been formed. After all, judging X wrong is supposed to be a practical matter; it is a
matter not only of what we think about X but what we do about X. That agiven
practice is to be toleratedthat it is not to be coercively suppressedseems
tomany people to represent a diminution of the practical force of the judgment that
it is wrong.
This view of the matter is best known from Susan Menduss book,
Toleration and the Limits of Liberalism, quoted by John Horton in his essay in
Heyds collection. Mendus writes: [W]here toleration is based on moral
disapproval, it implies that the thing tolerated is wrong and ought not to exist. The
question which then arises is why it should be thought good to tolerate. The
view is echoed in Bernard Williams contribution, Toleration: An Impossible
Virtue. Toleration appears impossible, said Williams, becauseit seemingly
required someone to think that a certain belief or practice was thoroughly wrong or
bad, and at the same time that there was some intrinsic good to be found in its
being allowed to flourish.
David Heyd in his Introduction to Toleration: An Elusive Virtue seems
sympathetic to the paradoxview,at least to the extent that he thinks that nothing
but a perceptual shifta change in ones way of seeing thingscan make sense
of toleration. I dont want to deny the interest of Heyds view: I think it is a most
welcome addition to our understanding of tolerancewhich is the ethical presence

1
I mean pragmatic, not in the sense of loosely consequentialist, but in the sense that is sometimes
contrasted with semantic and syntactic.
1
of something like toleration in personal life. 2 But I dont believe that this
perceptual shift is necessary in our understanding of toleration as a political
principle. (Heyd agrees: but thats because, unlike me, he doesnt think toleration
is a political principle.) I shall argue that at the political level there is no paradox
that needs displacing with a perceptual shift like this.
Nor, I want to suggest, is there any need for talk of a prima facie reason for
suppression being overbalanced by other stronger reasons having to do with
autonomy and respect for persons. That is not the logic of the situation. No doubt
if something thinks he does have reasons for suppressing, say, a religious practice,
we might want to put to him reasons concerning respect of or the autonomy of the
persons who would be the victims of such suppression. But the most important
part of the argument for toleration involves reminding the potential suppressor that
the judgment that a practice is wrong does not in itself generate even a prima
faciereason for suppressing that practice.
So: my aim in this paper is to argue that there is nothing paradoxical in the
concept of toleration. One can set up an appearance of paradox: thats what I did
in my first paragraph. But I believe the appearance of paradox is very easy to
dispel, and the dispelling of it is bound up with a perfectly sensible, broad,
coherent, and attractive understanding of what toleration involves. In short, I shall
argue that concern in the recent literature about the paradox of toleration is
artificial and somewhat over-wrought.

II
Dispelling the appearance of paradox can take either of two forms, each in its way
sufficient.
The first way of dispelling the appearance of paradox calls in question what
is seen (by those who talk of paradox) as the practical upshot of a moral judgment.
They say that a judgement that X is wrong provides a reason for suppressing X; I
shall show that this is mistaken.
The second way of dispelling the appearance of paradox calls in question the
philosophers insistence on focusing on a very narrow range of cases of pure
toleration and putting aside some evidently non-paradoxical casesthat they say
dont really involve toleration at all, despite being (in Brian Leiters phrase)
superficially similar to the central cases.Some of these superficially similar
cases involve a claim that the object of toleration ought not to have been the object
of a full-blooded moral assessment at all, certainly not by the entityusually the
stateon which the duty of toleration is normally said to be incumbent. I think

2
Unlike David (at p. 17n1 of the Elusive Virtue volume), I dont think toleration and tolerance are
interchangeable. The first is a political principle; the second is a personal virtue.
2
arguments of this kind are more central to the tolerationist tradition than many who
talk of paradox suppose.
Put this way, it looks as though the second way of dispelling the appearance
of paradox is prior to the first. We begin by questioning whether cases of
tolerating X necessarily presuppose a judgement that X is wrong. And then we
argue that, even if they do, such a judgement does not have any practical
implications that would conflict, even in a prima facie way, with what toleration
normally requires.But actually I want to consider matters the other way round. For
many of the arguments that are used to blunt the practical force of the judgment
that X is wrong also end up being arguments that show that certain sorts of
entitylike the statehave no business forming such judgments.
So, in what follows, I shall pursue the first way of dispelling the appearance
of paradox in section III and the second wayof dispelling the appearance of
paradox in sectionsIV and V.

III
The appearance of paradox turns crucially on a sense that a moral judgmentsuch
as that a given practice is wrongmust have certain practical consequences. It
must not be, so to speak, practically inert. The idea is that among the practical
consequences that such a judgment must have are actions that are prohibited by
toleration. For example, Bernard Williams implies that a strong condemnation of a
certain belief will normally generate a strong desire to suppress the belief in
question and drive it out. In his book Toleration:The Liberal Virtue (p.
9),Bicanahin suggests that moral values that lead one to condemn a practice also
strongly urge one to prohibit, hinder, or coercively interfere with it. Joseph Raz
seems to think that the moral disapproval presupposed by toleration generates an
inclination or desire to persecute, harass, harm or react in an unwelcome way to a
person who holds the beliefs or engages in the practices that are disapproved of.3
Toleration surely prohibits all these thingssuppressing other peoples beliefs and
driving them out, prohibiting, hindering, and coercively interfering with other
peoples religious practices, and persecuting, harassing, and harming those who
hold the beliefs or engage in the practices. But they are all said to be within the
range of the ordinary practical upshot of the judgment that the beliefs in question
are incorrect and the practices in question are wrong. It is this viewabout the
practical upshot of a judgment that something is wrongthat I want to consider.
Lets focus more closely on practical upshot. What is the connection
supposed to be? Three possible connections seem to be in play. (They are not

3
I use the term generates advisedly. For reasons I explain below, I think Raz denies that this is a matter
of logical implication.
3
mutually exclusive.) (A) There may be a psychological connection between
judging X wrong and doing or wanting to do something suppressive or coercive
about X. (B) There may be a logical or conceptual connection, whereby acting
against X seems part of what is (literally) implied by judging X to be wrong. Or
(C) the connection may be purely moral, mediated by an independently grounded
moral principle.

(A) Psychological connection. There may be a psychological connection

between judging X wrong and doing or wanting to do something suppressive or
coercive about X. This is undeniable. Taken on its own, however, the
psychological connection might be regarded by the proponent of toleration as fair
game for attack. Such attack will undermine the talk of paradox by undercutting its
premise.
After all, there are all sorts of psychological urges that we believe people
should abandon or suppress. For some unreconstructed male chauvinists, the
judgement that woman is beautiful generates an urge to whistle at her or harass her.
The connection is common and, in a sexist society, ordinary or normal. Members
of such a society might doubt the sincerity of Ps acknowledgement that a given
woman was beautiful if P was not prepared to act in that way. But there is no
paradox of sexual restraint in relation to peoples attempts to alter this. Feminists
and the men who support feminism are just trying to change what is the normal
course of behavior.
The same is true of supporters of toleration. They are trying to change what
for many people are or have been the psychological consequences of judging
others religious beliefs and practices wrong. Such judgements issue or used to
issue in a great deal of anger and hostility, as John Locke remarked, much more so
than in other areas of life:
In private domestic affairs, in the management of estates, in the conservation
of bodily health, ... [n]o man complains of the ill-management of his
neighbour's affairs. No man is angry with another for an error committed in
sowing his land or in marrying his daughter. ... But if any man do not
frequent the church, if he do not there conform his behaviour exactly to the
accustomed ceremonies, or if he brings not his children to be initiated in the
sacred mysteries of this or the other congregation, this immediately causes
an uproar.
On this account, the aim of the proponent of toleration is to break this emotional
link. This is something we do all the time, in processes of social, cultural, not to
mention legal change. Whats paradoxical about that?

4
 To put it another way: The problem is not in the idea of toleration. The
problem seems to be in the philosophers apparent endorsement of the urges and
inclinations that the tolerationist is trying to reform. The philosophers are
undermining the process of emotional reform by reinforcing the urges and
inclinations of those to whom the case for toleration is being addressed. They are
like chauvinists trying to undermine feminism. Dont worry about what John
Locke says about not getting angry at things you disapprove of, they say.
Getting angry at things you disapprove of is perfectly natural.
Of course this way of putting it is unfair (to the philosopher). For the
philosopher is not endorsing a psychological connection. The philosopher claims
to be pointing to a conceptual connection. He too may deplore the hostility that
emerges from adverse moral judgment. But he thinks the link between them is one
of implication not psychology. Endorseor deplore, it doesnt matter, says the
philosopher: its a matter of logic. You cant ask a person to give up the
implication of a judgement without giving up the judgment itself.

(B) Logical or conceptual connection. What might this alleged conceptual

connection between X is wrong and the appropriateness of suppressing X consist
in?
Well, there used to be a position called prescriptivism. It was key, for
example, to the meta-ethics of R.M. Hare (in The Language of Morals and in
Freedom and Reason). Prescriptivism took seriously the practical nature of moral
judgments, indeed value or normative judgments of any kind. Crudely understood,4
it held that calling something X involved issuing a prescription like Dont do X.
Now, such a prescription might be issued in the first instance to oneself.
Prescriptivists like Hare suggested that to assent to the judgment that X is wrong
is (among other things) to tell oneself not to do X, and the practical force of that is
that one prepares or intends not to do X, at least if some question of doing X has
arisen. If P says X is wrong and then promptly adopts the intention to do X, one
suspects P of insincerity in the judgment (absent some elaborate special
explanation). Part of the meaning of X is wrong is Dont do X (certainly as
addressed to oneself); and one does not form an intention to do X while addressing
Dont do X to oneself.
The prescription need not be confined to ones own choices and actions. For
P to tell Q that X is wrong isIm stating this in very crude termsfor P to tell
Q not to do X. With this, we seem to be in the ball-park so far as the practical
implications that are supposed to generate the paradox are concerned. A judgment
that X is wrong is not inert in respect of ones attitude towards others behavior.

4
I am leaving aside the universalizability element of Hares meta-ethics.

5
But still there is quite a distance to go before we get anywhere near the paradox of
toleration, because of course there is a considerable logical distance between
telling someone not to do X and calling for anything like the coercive enforcement
of that prescription.
For one thing, there are all sorts of problems with prescriptivism.I want to
mention only a few of these, and I am not going to rely on any of them (for my
argument against the claim that toleration is paradoxical). But prescriptivism does
have problems with the phenomenon of weakness of will; there are problems with
Hares suggestion that to call X wrong is to tell oneself not to do it; and there are
problems with his claim that to tell oneself not to do X is (more or less) the same
as resolving or intending not to do X. There are also problems with any attempt to
extend Hares prescriptive analysis to evaluative language in general. The position
looks attractive for straightforward normative claims like X ought to be done. It
doesnt look so convincing for Y is good, where Y is the name of a thing (such
as a car, a painting, or for that matter a person). Hares strategy was to parse Y is
good as implying Choose Y or If you are looking to choose something of the
kind to which Y belongs, then choose Y. But his prescriptivism looks a bit over-
extended at this point.
Never mind: suppose prescriptivism is true and that these problems can be
overcome. (If it is false, then I dont think the alleged appearance of paradox even
gets off the ground.) Still, we have not established anything remotely like a logical
implication between X is wrong and the kind of coercive actions in regard to X
that toleration is supposed to forbid. Certainly someone who says that X is wrong
must evince some sort of determination to suppress X so far as his own conduct is
concerned. But that is perfectly compatible with toleration. Someone who says
that X is wrong commits himself to refrain from X-ing. But there is no tension at
all between that commitment and his refusal to stop others from X-ing.
What about the other-regarding aspect of his prescription? If the
prescriptivist is right then saying in the presence of others that X is wrong seems to
involve telling them not to engage in X. Isnt that incompatible with tolerating
their X-ing? I think not. Terminology may confuse us here: telling someone not
to do something sounds a bit like commanding them not to do it, and if we make
that connection then we may be in danger of carrying over some sense of threat or
force often associated with command. But prescriptivism does not commit one to
saying that a judgment of wrongness has those overtones. Conveying any sense of
menace or coercion is something extra, and prescriptivism does not entitle us to
help ourselves to that as part of the logical analysis.
Even soyou may saysome vulnerable people may be coerced (in
some weak sense) just by anothers implying Dont do X when they say X is
wrong. We are back to psychology again, but no matter. If we expect this to be
6
so, then we might draw a different conclusion: from the fact that one judges X to
be wrong one cannot always infer that it is appropriate to say that X is wrong,
particularly when this would have to be said in the presence of someone who might
construe the saying of it as coercive. A tolerant attitude (broadly conceived) would
be willing to countenance this. And again there is no paradox here, because
nothing to the contrary is logically implied by ones subscription to the proposition
that X is wrong. One can subscribe to a proposition without being logically
committed to giving voice to it in the presence of others.
Logically, then, it is important to remind ourselves how limited the
implications are of a judgment that a belief is incorrect or a practice wrong. Even
if prescriptivism is true, the most that follows from such a judgment is ones
resolution to eschew such beliefs and practices and, under certain conditions, to
prescribe non-coercively to others that they should eschew them as well. For
normal cases, none of this comes anywhere near coercion. I believe that one of the
most important parts of a theory of toleration is to remind us of this.
That nothing whatever follows from X is wrong so far as the coercive
suppression of X is concerned is evident once we examine the logic of a normative
judgment and the logic of a coercive recommendation. The reasons that there must
be for saying X is to be coercively suppressed go way beyond the reasons that
would justify X is wrong. Acts of coercion raise issuessome of them
consequential issues, some of them issues of inherent moral significancethat
need to be dealt with on their own terms and that are not yet accounted for in the
judgment that the object of coercion is wrong. Here are some of those
distinguishing issues:
1. Coercion involves the deployment of individual and social resources,
and, if those resources are scarce, the wrongness of the intended
object of coercion does not begin to settle those issues. Many fewer
resources are involved in just making the moral judgment that X is
wrong.
2. Coercion involves a certain sort of impact on the life and agency of
those subject to it, which always goes beyond the mere evaporation of
the wrong action that is supposed to be its target.5
3. Coercion is experienced as demeaning and insulting (in a way that a
mere judgment voiced that ones conduct is wrong is not normally
experienced).

5
There is a fine account of this in Chapter 15 of Joseph Razs book, The Morality of Freedom, where Raz
insists (quite rightly) on a separation between (i) the implications that the wrongness of X may have for
any autonomy-based claim concerning the performance of X and (ii) the implications of any proposal to
coercively suppress or put a stop to X.
7
 4. Coercion is necessarily a sort of blunt instrument that will often chill
behavior adjacent to X even if that behaviour is not wrong like X is.
Moral judgments, by contrast, can be as fine-grained as our normative
language allows.
5. Coercion furnishes opportunities for various forms of corruption and
various forms of pathology on the part of the coercer. No doubt there
are some pathologies associated with the simple voicing of moral
judgements, but they are seldom as serious as those associated with
the use of coercive power.
6. Coercion often has an irreversible impact in the world, whereas
ordinary moral judgements, being revisable, do not. Of course one
can reverse or abandon a policy of coercion, but the effects of such
coercion as has taken place will remain.
There are probably lots of other features as well that distinguish coercion form
mere moral judgment. Since coercion has these features, we must expect the truth-
conditions of the claim that coercively suppressing X is appropriate to be quite
different from the truth-conditions of the claim that X is wrong. It is obvious then
that the former cannot be derived logically or conceptually from the latter.
What about the other way round? Maybe the judgment that X is to be
coercively suppressed implies that X is wrong. (I can think of various reasons for
denying this, but lets accept for the sake of argument that it is true). Then Xs
being wrong is a necessary condition for its being appropriate to coercively
suppress X. In this very minimal sense, the wrongness of X makes some
contribution to the case for coercive suppression. But the fulfilment of a necessary
condition is not usually understood as part of a justificatory argument. That I am
alive and breathing is a necessary condition of my being chosen as the next
astronaut to visit the international space station. But it does not really contribute
anything to the case for my being the next astronaut to visit the international space
station.
Lets sum up the argument so far. The beginning of wisdom in the
evaluation of anything that looks remotely like state moralism in relation to
religion or any other matter is the logical gap (even on a prescriptivist account)
between the practical implications of a judgment that X is wrong and the judgment
that X is to be coercively suppressed (or prohibited, hindered, harassed, etc.).
There is no logical or conceptual way of bridging that gap. And my point is that,
given the existence of this prescriptive gap, it is not at all clear how the so-called
paradox of toleration is supposed to arise.
Is any of this contested? Well, historically, some moral philosophers have
associated terms like duty and obligation with sanctions. John Stuart Mill

8
believed that the idea of a penal sanction, enters into the conception of any kind
of wrong. He said in Ch. 5 of Utilitarianism that
we do not call anything wrong, unless we mean to imply that a person ought
to be punished in some way or other for doing it. It is a part of the notion
of Duty, in every one of its forms, that a person may rightfully be compelled
to fulfil it.
I am not sure whether those who subscribe to theparadox-of-toleration thesis intend
to associate themselves with John Stuart Mill on this. Mills position is not now a
mainstream position. And anyway, Mills account of punishment and
compulsion was capacious enough to encompass non-legal and non-coercive
meanscompulsion if not by law, by the opinion of his fellow-creatures; if not
by opinion, by the reproaches of his own conscience.I think his position was that,
at a minimum, to judge something wrong is to anticipate that the wrongdoer will or
should be punished and constrained by his own conscience. And of course that
doesnt generate any sort of paradox so far as toleration is concerned. Beyond this
minimum, familiar doctrines associated with Mills own liberalism (e.g., in On
Liberty) entail a recognition, not a denial, of the gap between moral judgment and
the sort of coercive enforcement that toleration is usually said to be opposed to.

(C) Moral connection. Even if there is not a conceptual connection between

X is wrong and X is to be coercively suppressed, might there not be a moral
connection? Might not someone hold the following moral principle?
M: The coercive suppression of wrong action is usually
permitted (or required) just in virtue of its wrongness.
If M is a plausible principle, then the appearance of paradox revives. Though it is
not analytic, maybe M is nevertheless true or a moral commonplace. Arguing for
the toleration of X therefore does require us to come up with something that can
displace the force that the judgement that X is wrong normally has in the light of
M.
No doubt some people hold M. But once we acknowledge that it has to be
defended as a synthetic propositionthat it is definitely not a matter of the
prescriptive logic of a judgment that something is wrongit begins to look quite
implausible. This is particularly so for anyone who bears honestly in mind the six
distinguishing features of coercive suppression that we mentioned a page or two
earlier (pp. 7-8 above). Unfortunately the fact that a putative principle is, on
reflection, implausible does not mean that it is not widely held; people hold all
sorts of implausible and unpleasant moral propositions. An awful lot of the work
that has been done in the tradition of tolerationist arguments has involved arguing

9
against M. 6 Some of it involves highlighting and reminding people of
considerations 1 through 6 mentioned on pp. 7-8. Some of it also involves
reminding people of the futility and counter-productiveness of using coercive
means in support of moral judgments.
On both fronts, the arguments may be tailored particularly to religion.
Mostly in our discussion so far, I have imagined a simple judgement that a
practice, X, is wrong. But while everything said is applicable to the coercive
enforcement of straightforward moral judgments about wrong action, the
arguments are particularly powerful in the case of adverse religious judgments
about religious actions or practices. Of the considerations mentioned on pp. 7-8,
numbers 2 and 3 are particularly important for the religious case given the
prominence of religious practice in most peoples conception of themselves and of
what makes life worth living. And John Locke is famous for the view that, in the
area of religion especially, coercive means are by and large futile and non-
productive.7 Coercion cannot suppress conviction or elicit sincere belief; but the
attempt can still generate the bad consequences mentioned on p. 7.8
The important point to see is that these arguments against M are aimed at
cutting off or pre-empting anything that could possibly count as a paradox of
toleration. There is no paradox of toleration unless something like M can be

6
Our arguments against M will be rather like the arguments we use to try to break the spell of the
psychological association between moralism and coercion that I mentioned on pp. 4-5 above. The
difference is that the psychological association may be impervious to argument (and still be
objectionable).
7
Locke, however, was never able to make this stick as tightly as he wanted. Locke's criticsnotably
Jonas Proastargued persuasively that force may sometimes work indirectly to inculcate beliefs or to
make the mind receptive to beliefs, even if it does not work directly on the understanding. Locke
accepted this. There is no denying, he said, that in some particular cases force may work to procure
salvation. God himself makes use of all sorts of things to save men's souls (as our Savior did of clay and
spittle to cure blindness). God might have ordained the general use of force in religious matters, but he
didnt. So the position to which Locke retreats is that even though force may work in some cases, it has
not generally been ordained by God as a means of conversion: It is not for the magistrate, or any body
else, upon an imagination of its usefulness, to make use of any other means for the salvation of men's
souls than what the author and finisher of our faith hath directed. So the main line of Lockean argument
that survives Proasts critique is dominated by Lockes sense of what God has and has not ordained. (I
have adapted this material from pp. 210-11 of God, Locke and Equality. The relevant Lockean responses
to Proast are in theSecond, Third and (mercifully uncompleted) Fourth Letters Concerning Toleration.
8
This helps answer a criticism I made many years ago against Lockes argument. In Locke, Toleration,
and the Rationality of Persecution (reprinted in Liberal Rights), I complained that he showed only that
the enforcement of religious practice was futile not that it was wrong. But it is both: it is futile for just the
reasons that Locke insisted uponfire and the sword cannot produce sincere beliefand it is wrong
because the use of fire and the sword in a (futile) attempt to produce sincere belief always does great
moral harm to things like autonomy and is always an insulting violation of duties of respect owed to those
to whom the coercion is directed.
10
sustained. That, I guess, is why all this talk of a paradox of toleration strikes me as
so obtuse. Historically, many of the arguments brought forth in support of
toleration are arguments whose purpose and tendency is to dispel the main pillar by
which the alleged paradox is sustained.
One last objection: Isnt it a problem that in my arguments against M, and
more generally in my driving this huge wedge between X is wrong and X is to
be coercively suppressed, I am proving too much? For surely coercion in support
of morality is sometimes justified: we think this about the coercive upholding of
justice and basic rights, for example, or the coercive upholding of many of the
moral values protected by ordinary criminal law. How can this be, if M is always
false?
The answer is that M is not false for all cases. There is a class of cases
where it seems to be true, a class of cases captured by a variant of M, namely
MH: The coercive suppression of wrong action that causes harm
to others is usually permitted (or required) just in virtue of the
fact that it causes harm to others.
We can distinguish MH from M in several ways. Coercive interference can actually
have an effect in preventing harm; and the adverse impact on what seems to be a
persons autonomy can be justified by the need to protect others autonomy from
the harms that certain sorts of wrong-doing involve. The classic defenders of
toleration were well aware of this and they devoted a lot of time to showing that, in
the religious case, most beliefs and practices that were judged wrong were in fact
not harmful (except at most to the person engaging in them). Thats the force of
Lockes insistence, for example, that one man does not violate the right of another
by his erroneous opinions and undue manner of worship, nor is his perdition any
prejudice to another man's affairs. He wants to show that coercive interference
with religious practice and belief cannot be justified under MH.
Given the plausibility of MH, there might be something we could plausibly
call the paradox of toleration if the general tendency of tolerationist arguments
were to refrain from interfering with harmful religious practices. But that is not
and never has been the tendency of tolerationist arguments. On the contrary, the
defenders of toleration have always sought to limit the prohibition on coercive
interference at exactly the point at which religious practices become harmful.

IV
Now I want to consider the possibility of justifying toleration on the ground that
professional coercers (e.g. state officials) have no business judging religious
practices right or wrong (except on the grounds of their harmfulness).

11
 Most philosophers who talk of a paradox of toleration want to push this kind
of justification to one side. They want the starting point of any argument about
toleration to be a clear judgment about the wrongness of the practice, X, that is to
be tolerated. Thus the spurious connection we have been exploringbetween
judging X wrong and doing something coercive to suppress Xis given pride of
place in what we may call the narrow philosophical conception of toleration. They
want that connection there so that there can be a puzzle for them to work on.
These philosophers tell us that if a proposal to refrain from suppressing X is based
on the proposition that the potential suppressors really have no business forming a
judgment about Xs rightness or wrongness at all, then it is not really a proposal to
tolerate X. That X is to be tolerated is supposed to presuppose a judgement that X
is wrong, and wrong in a sense that has a conceptual link to the appropriateness of
suppressing X. Thats why the philosophers say that toleration properly
understood is paradoxical.
Now we have just seeninsection III (B)thatthere is no sense of
wrongness that is linked conceptually in this way to the appropriateness of
suppressing the practice supposed to be wrong. So there is no paradox, even if
tolerating X does presuppose Xs wrongness. But I dont think that even the
presupposition requirement can be defended. In this section I shall argue against
excluding, from the core range of toleration, arguments against coercion based
on the proposition that those in possession of the means of coercion should refrain
from forming views about what is right and wrong in matters of religion.
The case I shall make is partly affirmative, i.e. it is affirmatively in favour of
a broader conception of toleration. Toleration (especially religious toleration) has
been an important ideal in the liberal tradition since Lockes time, and I believe we
should understand it in a broad tradition-based way (just as we understand, say, the
rule of law or the separation of powers by reference to the tradition(s) in which
these ideals have been framed and argued about in the past 300 years or so). And it
is certainly the case, in the liberal tradition, that arguments based on the
inappropriateness of political authorities making moral judgements on religious
matters have played an important role in relation to the ideal of toleration. They
have not been the only arguments in favour of toleration, but they have stood
alongside some of the important arguments canvassed in section III. Not only that,
butand this is importanttheyrepresent positions that have, in a sense,
developed out of the arguments canvassed in section III.
I am heartened in this methodology by David Heyds observation in the
Nomos volume, Toleration and its Limits, p. 172, that

12
 [a]bstract theoretical analysis of the idea of toleration that ignores the way
the idea has operated in political rhetoric runs the risk of becoming
irrelevant.
Unfortunately no sooner did Heyd say this, than he went on to say that it was also
important to lter[ ] out those phenomena that do not satisfy certain theoretical
conditions. Heyd seems to believe that we owe something to both sides in this
methodological dispute. We owe tribute, on the one hand, to the historical
tradition in which arguments about toleration have been developed. And, on the
other hand, we owe tribute to the philosophers who have attempted to isolate a
particular narrow conception of toleration that, to them, is of surpassing theoretical
interest. I am sorry, but I just dont get the latter point. On what authority do the
philosophers insistas a theoretical conditionthat tolerating X has to imply or
presuppose that X is wrong (in some suitably robust sense of wrong)? After all,
this is not revealed by analysis of ordinary usage. It is, at best, something
constructed; and I think it is constructed just for the purposes of sponsoring
arguments of philosophical interest. The philosophers want to define a narrow
concept of toleration to give themselves an interesting conundrum to play with.
They actually want toleration to be a paradox, in regard to which they can then
show off their special skills and poise and clear-headedness. And it is pursuant to
this interestand this interest onlythat a narrow definition seems desirable.9 As
Heyd puts it in his introduction (p. 5) to the Elusive Virtue collection, It is an
interesting feature in the analysis of toleration that the narrower its definition the
more paradoxical it becomes.
As the tone of these comments suggests, I think there is also a negative case
to be made against the philosophers narrow definition. On the one hand, they have
no right, so to speak, to narrow the discourse of toleration in this way. And
secondly, the point of their doing so fails anyway. For if what we just said in
section III is correct, you are not going to get a paradox to play with even if you do
confine toleration to contexts where the thing to be tolerated is judged morally
wrong. So even the insiders professional-entertainment case for narrowing the
definition fails to get off the ground. We might as well stick with a broader
conception, since narrowing our understanding of toleration does not promise the
intellectual fun that the philosophers were hoping for.
Is this too quick? One response might be: Well, what if there is someone
some potentially intolerant ruler, for examplewhojudges some religious practice

9
So, on my view, philosophers should not be bullying their audience by representing a failure to infer X
is wrong from X is to be tolerated as some sort of intellectual or conceptual mistake. The mistake, if
there is one, is in failing to go along with the philosophers sense of what makes an argument about
toleration philosophically worth spending time on.
13
X to be unequivocally wrong? Dont we want to have a tolerationist argument
against him, whatever other arguments we have going in this broad area? Well,
yes we do. But it need not be a tolerationist argument that leaves his judgment that
X is wrong entirely untouched or unquestioned. If we do leave it untouched or
unquestioned, he wont get very far with it; thats what we showed in section III.
But we may not be willing to let him get even that far. We may want to show that
intellectually or politically, he is not entitled to form a judgment about Xs
wrongness. And indeed we may think that this is where the action is or where the
action ought to be so far as the case for toleration is concerned. We may think that
if we cant make the case here, then we have failed.
Our interlocutor may persist: Yes, but what if we cant shake the ruler free
of the judgment that X is really wrong and that this is something he is entitled to
form a view on? Well, in the face of this possibility, we may comfort ourselves
with the thought (based on what we said in section III) that the ruler still has a long
way to go before he can infer the appropriateness of coercion from whatever he
insists on saying about Xs wrongness. But we may think still that there is a pro
tanto failure in the case for toleration if things get that far. Or, more moderately,
we may think that both phases of the argument are important, and that it is quite
misleading to say that, strictly speaking, only the second phase is an argument
about toleration.

V
David Heyd (Is Toleration a Political Virtue?) thinks that toleration cannot really
be regarded as a principle of the political philosophy of the liberal state. This is
not of course because he thinks such states are or should be intolerant. It is
because he thinks there is no room for the question of toleration to even arise:
Unlike a medieval sovereign, the state is an impersonal institution which
cannot be described as suffering in having to reconcile itself with beliefs
and practices to which it does not subscribe. Hence, it cannot be said to
overcome or endure its wish to undermine or interfere with them. In other
words, the state cannot be engaged in toleration.
His view is that the state cannot judge that a religious practice X is wrong. It has
no authority, means, or jurisdiction to do so. It follows, Heyd thinks, that the state
cannot be said to do anything or take any position that presupposes a judgment that
X is wrong. Buton the narrow conceptionthe proposition that X is to be
tolerated does presuppose that X is wrong. Accordingly, on Heyds view, the state
cannot hold that X is to be tolerated.

14
 I think this is all too quick. It fails to credit the role of tolerationist principles
in bringing us to the position (about the state) that Heyd simply announces. Let me
explain.
I assume that Heyd has no wish to deny that a liberal state may be, so to
speak, beset by judgments that X is wrong (beset by such judgements, surrounded
by them, pressured by them, and so on). A liberal state is open to (or at least hears)
the views and demands of its citizens, even views and demands that its liberal
character makes it incapable of acting on. Moreover, the liberal state must be
exhibit a certain resiliencein the face of such demands. It must be able say to its
citizens: These are not views that we (the state) may hold or act on. And if the
citizens ask Why not? the state must have an answer to give
I believe that the answer to be givenor some of the various different
answers that it is appropriate for the state to giveare either tolerationist answers
or are constructed out of tolerationist answers. We can miss this if we see the
modernity or liberalism of the state as something simply given: e.g. the state just is
an impersonal institution that is incapable of holding views about the wrongness of
religious practices. But that is not just given. Historically this characterization of
the entity that monopolizes force in society had to be constructed, and that
construction was partly argumentative; and, even in the present, the liberal, neutral
and impersonal character of the state has to be continually explained and defended.
When I first wrote about John Locke on toleration many years ago (in
Locke, Toleration, and the Rationality of Persecution, in Liberal Rights), I found
a number of commentators saying that Lockes position was based on the premise
that religion was not part of the function of the state. I wrote that that couldnt be
his premise: it would just be question-begging. What Locke had to do was argue
that religion was not part of the function of the state. And this, I said, he did do;
and the arguments he made remain important and available to modern liberals in
explaining and defending the impersonal neutral character of the state that they
have helped bring into existence.
Many of those arguments are versions of those we already considered in
section III. For example, Locke argued again and again in the Letter Concerning
Toleration that coercive means could do no real good in respect of religious
practice and belief. (I wont repeat the arguments here; they are very well known,
and I have written extensively about them elsewhere.) These arguments were
addressed potentially to anyone with coercive means at their disposal. For
example, they were part of the basis on which Locke established that it was not
appropriate for bishops and other religious leaders to use force against members of
their congregations. They were also part of the basis on which Locke argued
against the use of force by private individuals against others who differed from
them in matters of religion.
15
 But the futility-of-the-use-of-force argument crystallized in a particularly
powerful way in the case of magistrates, because of the way Locke defined
thestate. The magistrates power consists only in outward forcea magisterial
care ... consists in prescribing by laws and compelling by punishments. If we
define the state in this way, then it follows that it cannot have the enforcement of
true religious practice and belief as its object. This argument is not based on a
functional assumption. It is based on a modal assumptionabout the means that
states characteristically useand it works from there to a functional conclusion.
And just because the state is (by definition) a mechanism of this coercive kind, it
has no business forming judgments about matters on which coercion is inherently
inappropriate.
For this reason, I think it is very artificial to sideline the liberal-state
argument to the margins of tolerationist discourse. We cannot understand the
liberal-state argument without understanding the classic case for toleration in
which it is rooted. And, as I have indicated already, the philosophers attempt to
sideline this argument because it fails to give them the sort of paradoxical
entertainment they wantwhich they cant have anywayis particularly silly.

VI
One last thing.I have heard it said that there is anotherand perhaps more
respectablereason for distinguishing arguments for toleration from arguments
about the inherent inappropriateness of the modern liberal state making judgments
about religion. The latter arguments, it is said, are much more respectful of
religious minorities than toleration arguments. Toleration arguments always seem
to be condescending, judgmental, and insulting. The tolerator says,
T1: I believe that what you are doing is wrong, wrong, wrong, but as a matter
of grace and favour (or for reasons I can barely formulate) I will tolerate
your practice, that is, I will permit you to do it anyway, despite my strong
and justified inclinations to the contrary.
The modern liberal, by contrast, just says: You have a right to religious freedom,
and the character of your beliefs is simply none of my business. That is much less
unpleasant, much more respectful.
So stated, we can agree with the assessment. But again we should dispute
both the characterization of the tolerationist attitude and the wedge that is being
driven between that attitude and the position of the rights-respecting state. The
characterization of the tolerators position in the last paragraph depends on
something like the paradox of toleration. If that talk of paradox makes no sense
and that is what I argued in section IIIthen we are not entitled to present the

16
tolerator in this light, and the tolerator, if he knows his business will not describe
his position in that way. He may say instead:
T2: I disagree with your practice and, as you know, I judge it wrong. But I
acknowledge that that does not in any way make it appropriate for me to
interfere with what you are doing, especially since it is not harmful.
Not only that, but if he is a magistrate, he will add:
T3: Moreover, in my capacity as magistrate (that is, as someone entrusted
specifically with the coercive power of the community), I want to say that it
is not for me to form any judgement whatever about your practice. If you
want to have a debate about it, off duty and in private life, thats fine. But in
my official capacity, all I can say is that any coercive interference with what
you are doing wouldbe wrong.
Together, T2 and T3 define an attractive and respectful position and one that has the
advantage of making the argument for the modern rights-based view clear and
transparent.
It seems to me to be an enormous advantage of my rejection of all this talk
about a paradox of toleration that I can put T1 aside as a characterization of the
tolerationist attitude and replace it by the much more attractive, much more
illuminating, and much better-thought-through version in T2 and T3.

17